passenger car equivalents of trucks under lane restriction
TRANSCRIPT
Louisiana State UniversityLSU Digital Commons
LSU Master's Theses Graduate School
2009
Passenger car equivalents of trucks under lanerestriction and differential speed limit policies onfour lane freewaysJohn StanleyLouisiana State University and Agricultural and Mechanical College, [email protected]
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_theses
Part of the Civil and Environmental Engineering Commons
This Thesis is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSUMaster's Theses by an authorized graduate school editor of LSU Digital Commons. For more information, please contact [email protected].
Recommended CitationStanley, John, "Passenger car equivalents of trucks under lane restriction and differential speed limit policies on four lane freeways"(2009). LSU Master's Theses. 3959.https://digitalcommons.lsu.edu/gradschool_theses/3959
PASSENGER CAR EQUIVALENTS OF TRUCKS UNDER LANE
RESTRICTION AND DIFFERENTIAL SPEED LIMIT POLICIES ON
FOUR LANE FREEWAYS
A Thesis
Submitted to the Graduate Faculty of the
Louisiana State University and
Agricultural and Mechanical College
in partial fulfillment of the
requirements for the degree of
Master of Science in Civil Engineering
in
The Department of Civil and Environmental Engineering
By
John Stanley
B.Tech, University of Kerala, 2004
May, 2009
ii
ACKNOWLEDGEMENTS
First and foremost, I would like to give my deepest and sincerest thanks to my advisor, Dr.
Sherif Ishak, for his brilliant guidance, encouragement and support and funding me throughout
the program. Thank you for always being there to help me and support my study. My heartfelt
gratitude to Dr. Clifford Mugnier, who has been my guide as well as a guiding mentor to me. I
would like to give my special thanks to Ciprian for his invaluable help and suggestions. Thank
you for your unselfish and timely help.
I also want to acknowledge my colleagues, Cherian, Pradeep, Bharath, Chaitanya, Mini, for
their precious assistance. I do wish to thank Hannah, Wakeel and Ravi for their constant support
and camaraderie in the Lab. Thanks are due also to all my friends, Yoseph Marcos, Heui Yang
and Phillip Ebenezer and all in the International ministry team for their love and help in time of
need. Thank you for listening to me and for your words of encouragement. Your support meant
the world to me.
I would like to thank my dear parents who saw the ups and downs in my life, the tragedies
and triumphs, and still motivated me to go on and give back something to the world. I would
also want to thank my extended family in the U.S who helped me to blend in with the culture and
encourage my endeavor here.
Finally and most importantly, I thank my Lord Jesus Christ, without whom I consider myself
nothing and void. I am what I am now just by His grace.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ............................................................................................................ ii
LIST OF TABLES ...........................................................................................................................v
LIST OF FIGURES ....................................................................................................................... vi
ABSTRACT .................................................................................................................................. vii
1. INTRODUCTION .......................................................................................................................1
1.1 INTRODUCTION AND PROBLEM STATEMENT .......................................................... 1
1.2 RESEARCH OBJECTIVES………………………………………………………………..2
1.3 SCOPE OF STUDY .............................................................................................................. 3
2. BACKGROUND .........................................................................................................................4
2.1 LANE RESTRICTION FOR TRUCKS ............................................................................... 4
2.2 DIFFERENTIAL SPEED LIMIT (DSL) .............................................................................. 8
2.3 PCE RELATED STUDIES................................................................................................. 11
2.4 SUMMARY ........................................................................................................................ 13
3. STUDY SECTION AND DATA COLLECTION ....................................................................14
3.1 STUDY SECTION ............................................................................................................. 14
3.2 DATA COLLECTION ....................................................................................................... 14
3.2.1 DATA COLLECTION EQUIPMENT - RTMS .......................................................... 15
3.2.2 DATA COLLECTED .................................................................................................. 15
4. METHODOLOGY ....................................................................................................................17
4.1 INTRODUCTION .............................................................................................................. 17
4.2 CODING AND INPUT FOR SIMULATION .................................................................... 17
4.3 INPUT/ CONTROL VARIABLES .................................................................................... 18
4.3.1 SPEED DISTRIBUTIONS .......................................................................................... 18
4.3.2 VEHICLE TYPES ....................................................................................................... 19
4.3.3 PERCENTAGE OF TRUCKS IN THE TRAFFIC MIX ............................................ 21
4.3.4 RATE OF COMPLIANCE TO THE RESTRICTION POLICIES ............................. 21
4.4 SIMULATION CASE SCENARIOS ................................................................................. 21
4.5 ESTIMATING ET USING SIMULATION ........................................................................ 23
4.5.1 FREEWAY CODING IN VISSIM .............................................................................. 23
4.5.2 CALIBRATION AND VERIFICATION .................................................................... 23
4.5.3 ESTIMATION OF ET FROM SIMULATION OUTPUT ........................................... 24
5. SIMULATION RESULTS ........................................................................................................26
5.1 WHEN NO GRADIENTS ARE TAKEN INTO CONSIDERATION............................... 26
5.1.1 TRUCK LANE RESTRICTION ONLY ..................................................................... 27
5.1.2 COMBINED EFFECT OF BOTH RESTRICTION POLICIES. ................................ 29
5.2 EFFECT OF RESTRICTIONS WHEN GRADES ARE APPLIED .................................. 32
iv
6. CONCLUSIONS........................................................................................................................34
6.1 STUDY CONCLUSIONS AND FUTURE WORK ........................................................... 34
REFERENCES ..............................................................................................................................36
VITA ..............................................................................................................................................38
v
LIST OF TABLES
TABLE 1: VEHICLE CLASSIFICATIONS ............................................................................................ 15
TABLE 2: DESCRIPTION OF DATA COLLECTED BY RTMS ............................................................... 16
TABLE 3: DISTRIBUTION OF DESIRED SPEEDS FOR PASSENGER CARS (PC) AND TRUCKS ................. 22
TABLE 4: GRADES AND THEIR RESPECTIVE LENGTHS FOR EACH SCENARIO .................................... 22
TABLE 5: CASE SCENARIOS FOR EACH SPEED DISTRIBUTION .......................................................... 23
TABLE 6: SIMULATION SCENARIOS ................................................................................................. 26
TABLE 7: T-TEST FOR SIGNIFICANT DIFFERENCE OF ET UNDER TRUCK LANE RESTRICTION ONLY ... 28
TABLE 8: APPROXIMATION OF ET UNDER TRUCK LANE COMPLIANCE RATES .................................. 29
TABLE 9: T-TEST FOR ET UNDER TRUCK LANE RESTRICTION AND DIFFERENTIAL SPEED LIMIT ....... 31
TABLE 10: APPROXIMATION OF ET FOR DIFFERENT COMPLIANCE RATES TO DIFFERENTIAL SPEED
LIMITS .................................................................................................................................... 31
TABLE 11: SIGNIFICANCE OF ET ON UPGRADES COMPARED TO HCM ............................................ 33
vi
LIST OF FIGURES
FIGURE 1: TRUCK LANE RESTRICTION .............................................................................................. 2
FIGURE 2: DIFFERENTIAL SPEED LIMIT ............................................................................................. 2
FIGURE 3: STUDY SECTION WHERE DATA COLLECTED ARE MENTIONED AS SITE # ......................... 14
FIGURE 4: SPEED DISTRIBUTION CURVES FOR PASSENGER CARS FROM FIELD DATA ...................... 19
FIGURE 5: SPEED DISTRIBUTION CURVES FOR TRUCKS OBTAINED FROM FIELD DATA .................... 19
FIGURE 6: SPEED DISTRIBUTION CURVE IN VISSIM, FROM FIELD DATA ......................................... 20
FIGURE 7: BASIC VEHICLE TYPES USED FOR THIS RESEARCH ......................................................... 20
FIGURE 8: EFFECT OF % TRUCKS ON ET UNDER VARIOUS COMPLIANCE RATES TO TRUCK LANE
RESTRICTION. ......................................................................................................................... 28
FIGURE 9: PERCENTAGE OF TRUCKS PLOTTED AGAINST ET FOR 100% COMPLIANCE ...................... 30
FIGURE 10: PERCENTAGE OF TRUCKS PLOTTED AGAINST ET FOR 75% COMPLIANCE ...................... 30
FIGURE 11: PERCENTAGE OF TRUCKS PLOTTED AGAINST ET FOR 50% COMPLIANCE ...................... 30
vii
ABSTRACT
Lane restriction for trucks and differential speed limits for trucks and cars are becoming
more common and feasible policies to improve the efficiency and safety of a freeway. It is
believed that passenger car equivalents for trucks are impacted by these non typical freeway
operating conditions, which are not explicitly addressed by the latest edition of the Highway
Capacity Manual(HCM). Using simulated and real world data an elevated 18- mile four lane
freeway was modeled under the restriction policies. The section which was used as a test bed
was simulated under various control variables. Some of the control variables used were speed
distributions from the field data, truck percentages in the traffic mix and the compliance rate to
the restriction policies. The simulated results were compared with the corresponding values in
HCM and observations were made which can be used for further research. The simulated results
show that the ET values decreases with increase in truck percentages under the influence of the
truck restrictions due to “platooning effect” caused due to increase in the truck percentage.
1
1. INTRODUCTION
1.1 INTRODUCTION AND PROBLEM STATEMENT
As a result of the safety concerns associated with the ever-increasing number of annual
vehicle-miles traveled by passenger cars and trucks on freeways, several states have adopted
truck lane restriction and differential speed limit policies on specific freeway segments to
improve freeway operation and safety. Several studies have been conducted over the past few
years to evaluate the operational and safety benefits of such policies on trucks and traffic in
general. Most studies focused on policies implemented on freeway segments with at least three
lanes in each direction.
In the state of Louisiana, the same policies were implemented over an 18-mile elevated rural
section of Interstate 10 (I-10) in response to an 11-vehicle crash caused by a truck failing to
notice stationary traffic ahead in September 2003. The crash resulted in five fatalities and
prompted the state DOT to restrict trucks to the right lane (Figure 1) and reduce their speed limit
to 55 mph while allowing a speed limit of 60 mph for cars (Figure 2).
While similar policies were implemented in other states (e.g. Texas and Tennessee), the
literature review showed that truck lane restriction and differential speed limits had not been
implemented or evaluated on freeway segments with only two lanes in each direction. The focus
of this study is on evaluating the impact of trucks on traffic flow under both lane restriction and
differential speed limit policies. The effect is measured in terms of the passenger car equivalent
for trucks ( ) under a wide range of operating conditions that are not currently covered in the
latest version of the Highway Capacity Manual (HCM). is typically required for operational
and design analyses of basic freeway segments. The U.S. HCM provides estimates for freeway
capacity that are calibrated to a set of ideal conditions. Among those ideal conditions is the
stipulation that the traffic stream is uniform and consists of passenger cars only. In most
instances, prevailing conditions are not ideal and the traffic stream usually contains a mix of
different vehicles, i.e. trucks, buses, RV’s, and passenger cars. The HCM capacity analysis
procedures utilize Passenger Car Equivalents (PCE’s) to account for the presence of heavy trucks
2
Figure 1: Truck lane restriction
Figure 2: Differential speed limit
in the traffic stream. Using those PCEs, a non-homogeneous mix of vehicles in a traffic stream
can be expressed in a standardized unit of traffic. Though essential in carrying out capacity
analyses, those PCEs have been the subject of an old and long argument PCEs, a non-
homogeneous mix of vehicles in a traffic stream can be expressed in a standardized unit of
traffic.
1.2 RESEARCH OBJECTIVES
The primary focus of this research is on evaluating the impact of trucks on traffic flow under
both lane restriction and differential speed limit policies using VISSIM as a simulation tool.
3
More specifically, the research objectives of this study are to:
1. Quantify the impact of these restriction policies on the passenger car equivalents of
trucks (PCE’s or ET) using VISSIM.
2. Compile and compare PCE’s obtained from simulation results with HCM, assuming
different gradients and their respective lengths in the traffic stream.
3. Examine a hypothesis that the PCE’s provided in the HCM 2000 underestimates the
effect of heavy vehicles on the freeway under different operating conditions and
restriction policies.
1.3 SCOPE OF STUDY
The scope of this research study is limited to the Atchafalaya Basin section of I-10, which is
one of the elevated sections of the freeway where the new policies of truck lane restriction and
speed limit differentials were implemented. For the purpose of this study, traffic data was
collected from specific locations on the freeway segment. The collected data were used to study
the impact of the restriction policies and also as input data for the simulation tool.
4
2. BACKGROUND
As truck volumes continues to be on the rise on U.S. freeways, both elected officials and the
general public have suggested the use of restrictions policies such as to restrict the trucks to just
one lane and also to adopt differential speed limits for passenger cars with respect to trucks.
Each of these policies may be imposed on a section of a roadway either separately or together in
a combined strategy to get the desired results and improved safety and efficiency. This chapter
explains the prior research done to study the impact these policies on the overall operational
efficiency and safety. The studies conducted to quantify the impact on the traffic stream in terms
of passenger car equivalents of trucks(PCE’s) forms the closing section of this chapter. The need
for this research study resulting from inadequacies of previous studies, is also presented.
2.1 LANE RESTRICTION FOR TRUCKS
Lane restriction strategy for trucks is implemented to restrict trucks to a certain lane or lanes
and minimize the interaction between trucks and smaller vehicles. Since the traffic and highway
geometric conditions are different, there have been several possible design alternatives for lane
restriction for trucks. Researchers collected field data and/or simulated the traffic operation on
the road to investigate the impacts of lane restrictions. Only a limited number of safety related
studies exist in the literature, of which several analyzed efficiency of lane restriction and speed
limits in confined roadway conditions (i.e. overpasses, long bridges and super-elevated ramps).
Zavoina et al. (2) conducted a study to evaluate the operations of truck restrictions on I-20
near Fort Worth, Texas. The restrictions on the specified Interstate section were prohibiting
trucks from traveling in the left lane on a three-lane section. The study section didn’t have any
exit ramps.The study concluded that although the directional distribution of trucks changed
significantly due to the imposed restriction, no effects have been identified that could be
attributed to the truck restriction in the directional distribution of cars, speed of either cars or
trucks, or time headways between vehicles.
A report by Hoel and Peek (3) investigated the impacts of lane restriction on traffic flow
5
elements such as density, lane changing, and speed variance. Three sites were chosen on I-81 in
Virginia. The data was collected for four initial volume distributions on 0%, 2%, and 4% grades
at these sites. FRESIM simulation model was used to approximate traffic flow elements. Two
different restriction strategies were tested: restricting trucks from the left lane and restricting
trucks from the right lane. The authors recommended restricting trucks from using the left lane
on grades 4% or steeper.
Mussa and Price (4) examined the safety and operations on I-75 in Florida, where a median-
lane restriction for trucks takes place. Their particular objective was to find out the influence of
the restriction on truck operating speed and travel time throughout the day based on field data
and simulation (i.e. CORSIM 5) as well as crash data. The authors’ findings revealed that the
current policy of restricting trucks from the median-lane provided safety and efficient operation,
and therefore should be left in place.
Several other studies that looked into truck lane use restrictions found that there a
considerable improvement of safety conditions, but there is no significant impact on freeway
operations (see, for example, Borchardt (5) and Zeitz (6))
Models were developed by Gan and Jo (7) to find out the strategy for truck lane restrictions
that offered the most efficient operations on highways. The performance criteria included
average speed, throughput, speed differentials, and lane changes. Number of lanes, interchange
density, free-flow speeds, volumes, truck percentages, and ramp volumes were given. The
simulation results showed that average speed increased when the interchange density, truck
volume, and ramp volume were low. Throughput increased when the number of restricted lanes
increased. Low number of restricted lanes (ex: one out of three) brought higher capacity than the
non-restriction case for maximum truck percentage of 25%. The authors concluded that in
general, when the section with restricted lanes is not under heavy weaving and lane changing
conditions such as sections with densely spaced interchanges, having restricted truck lanes is
beneficial operational wise. On the other hand, there was considerable speed differential
between restricted and non-restricted lane groups, and the magnitude increased proportionally
6
with the increase in the number of interchanges, ramp volumes, truck percentages, and free-flow
speed. Another important point was that truck lane restrictions decreased the frequency of traffic
accidents by separating the slower vehicles from the faster ones and reducing the frequency of
lane changes. The appropriate number of lanes to be restricted was stated as one lane on three-,
four-, five-lane highways, and two lanes on four-, and five-lane highways if the interchange
density was not high and the truck percentage was at or below the average.
Harwood et al. (8), synthesized the knowledge about the safety interaction of trucks and
buses with the highway elements, and then analyzed the assembled information. They also
determined what could be done to improve the heavy vehicle safety on roadways. The results of
the study revealed that the fraction of the highway agencies that used or were considering the use
of differential speed limits was only 40%. However, the safety benefits of differential speed
limits were not proven. In fact, the study suggested that the speed variance between the
passenger vehicles and heavy vehicles might cause more traffic accidents. Truck lane
restrictions, according to this study, did not demonstrate any safety benefits nor did it show any
negative impact on highway safety in most of the evaluations. On the contrary, a recent test in
Houston, which lasted eight months, reported a safety benefit of restricting trucks from using the
left lane. Harwood et al. recommended conducting more research on differential speed limits
and truck lane restrictions in order to find out their impacts on highway safety.
A simulation study by Cate and Urbanik (9) showed the effect of prohibiting trucks in the
left lane on 3-lane highways. The VISSIM traffic simulation model was used to test different
scenarios and analyze the results. Truck lane restriction caused a slight increase in the traffic
density and level of service on flat grades. However, as upgrades approached 4%, the impact
became more significant. Similarly, the average travel time was affected slightly on flat grades,
although it reduced considerably on steep (> 4%) upgrades.
The study also showed that speed differential between cars and trucks was less than 1 mph
on flat sections, while it climbed up to 9.9 mph on steeper sections of the highway. Another
variable tested was the occurrence of lane changing. The reduction in lane changing behavior by
7
trucks surpassed that by cars on flat sections, but they were almost the same on upgrade sections.
The safety problem generated from the speed differential between cars and trucks was offset by
the safety benefits of reduced lane changing. Overall, prohibiting trucks from using the leftmost
lane on highways with three or more lanes in the same direction had no negative effect on
highway safety or operating efficiency.
Cate et al. (10) specified the impacts of lane use restrictions employed for large trucks on
Tennessee’s highways, and set guidelines for implementing these restrictions after a thorough
observation of lane use restriction practices in other southeastern states. Tests showed that even
with minimal use of signage and enforcement, the truck percentage in the left lane decreased
significantly after the lane use restriction was put into practice. The study recommended that
truck lane use restrictions be applied on freeways with at least 3 lanes in one direction. Also,in
this study which concentrated on safety issues, restricting trucks to a single lane was not advised,
because the barrier effect and the accelerated pavement wear it might cause would prevail over
the potential benefits of the restriction.
Although truck speeds increased in a few observations, the study showed a slight decrease in
its measure. Truck speeds being higher than the posted speed limit could be argued for its safety
benefits. Overall, lane use restrictions provided few tangible operational and safety benefits, and
produced the insight of enhanced safety and comfort for the majority of motorists. After meeting
all other requirements, the public insight would help the widespread practice of the truck lane
restrictions in Tennessee.
A study by Knipling et al. (11) stated that the purpose for implementing lane restrictions had
more to do with improving efficiency of a freeway rather than enhancing safety. In fact, it was
mentioned that lane restrictions, in some cases, created unfavorable effects on highway safety.
Knipling et al. pointed out that the truck lane use restriction was appropriate for interstates with
at least three lanes in one direction. Issues involved with implementing a lane use restriction
strategy were detailed in the report. Speed differentials and lane changes were considered
substitutes for safety measures, and some of the factors causing the accidents. To ensure the
8
safety benefits of potential lane restrictions, it was recommended that pilot studies be
implemented. Most of the time, lane use restrictions required the authorization of the legislation,
but the legislation sometimes authorized state DOT’s and local agencies to apply the lane use
restrictions on facilities under their control. All the stakeholders, primarily law enforcement
officials and organizations that represent the commercial transporters using heavy trucks, should
be included in the implementation stage of the lane use restriction. Awareness of the views and
perceptions of the truck operators were essential to the success of the lane use restriction
program.
Hanscom (12) compared truck lane restriction on a three-lane road with truck lane restriction
on a two-lane road. The three-lane section is located in an urban area near Chicago and the
trucks were prohibited to travel in the most left lane, while the two-lane site is a rural Interstate
section in Wisconsin with pavement deterioration, which prevents the trucks from traveling in
the right lane. It is concluded that beneficial traffic flow effects (e.g. reduced congestion) are
associated with the left-lane truck restrictions on three-lane roadways. On the other hand, the
author’s findings on the two-lane restrictions site include high violation rates and slowing of
impeded vehicles which raise safety issues.
Overall, the past research seems to indicate very little concrete evidence that operations or
safety are improved by the use of truck lane restrictions in two-lane highways. One key area of
agreement is that the use of truck lane restrictions on extended upgrades does improve operations
by reducing density and the number of lane changes, though safety may be compromised by the
resulting increase in speed differential between cars and trucks. However, surveys of motorists in
Texas and Washington State do show that motorist comfort is significantly increased in the
presence of truck lane restrictions.
2.2 DIFFERENTIAL SPEED LIMIT (DSL)
Differential speed limit is to set lower speed limits for trucks, compensating for their differences
in operational characteristics. Limited literatures on DSL were available compared to studies on
9
truck-lane restrictions. Most DSL studies focused on the impact of speed variance on traffic
accidents.
Wilmot and Khanal (13) reviewed the effects of speed limits on vehicle speed and safety on
roadways. They stated that there was no proof of the positive impact of differential speed limits
on highway safety. In addition, the difficulty of differentiating day and night at dawn or dusk
was the shortcoming of employing differential speed limits based on the time of day. Besides,
applying differential speed limits at urban boundaries created a problem: renewing the start and
end of differential speed limits when urban areas grew rapidly. Additionally, using differential
speed limits required extra signs and supplementary enforcement, raising cost issues.
A survey conducted by Sunbelt Research Corporation (14) questioned Louisiana drivers
about the 55 mph speed limit and other highway safety issues. The survey results showed that
most of the motorists drove faster than the speed limit on interstate highways. The respondents
who were not in favor of the speed limit change formed two thirds of the interviewees. The
majority of the drivers who often exceeded the speed limit were a part of this fraction of
interviewees. Half of those who thought the speed limit should change stated that 60 mph was a
reasonable speed limit. According to most of the respondents, the reasons for those who thought
speed limit enforcement was performed by the state police formed half of the interviewees; those
who believed it was not supposed or it was unpredictable comprised of the rest. Majority of the
respondents saw enforcement as an essential factor for higher compliance rates, while only one
fifth claimed education and advertising would be the solutions.
Monsere et al. (15) evaluated the effects of a proposed maximum speed limit change to 65
mph for trucks and 70 mph for passenger cars on Oregon’s interstate highways. The maximum
posted speed was 55 mph for trucks and 65 mph for passenger vehicles at the time of the study.
The report examined the influences of speed change on motor-vehicle accidents, enforcement,
health, economy, and the environment. The results indicated negative effect on all but travel time
and some economic development benefits.
10
A report, by Garber et al. (16), judged the safety effects of differential speed limits on rural
interstate highways against those of uniform speed limits. It was found that changing from a
uniform speed limit to a differential speed limit or vice versa had no impact on the mean speed
and speed variance of vehicles on highways. Also, crash rates had no association with the type of
speed limit chosen.
Milliken et al. (17) reviewed the current practice for setting and enforcing speed limits on all
roads in the United States and guided the state and local officials on a suitable technique to set
and enforce speed limits. Milliken et al. stated that there was a tradeoff between safety, travel
efficiency, and rationality of enforcement when speed limits were being set. Most of the time,
safety became the determining factor, because severity of traffic accidents depended on the pre-
crash speed of the vehicle. Higher speed limits caused increases in the speed dispersion. The
higher the speed dispersion on rural Interstates, the more crash fatalities there were. The
minimum speed dispersion was obtained when there was 5-10 mph difference between the road
design speed and the posted speed.
Another factor that triggered crashes was the great difference in the speeds of the vehicles
on a portion of the highway. This was seen in the area around interchanges. In fact, the high
traffic volume near interchanges on urban interstates increased crash rates, which indicated the
role of traffic density in the occurrence of traffic accidents. Milliken et al. noted that
enforcement and creative engineering measures were necessary for desired driver compliance
with the posted speed limits.
Considering these shortcomings, more appropriate actions, where necessary, are
implemented instead of applying differential speed limits to the entire network. The possible
actions are speed zoning at sites where lower speed was warranted and situating warning or
regulatory signs dedicated to trucks in order to differentially control their speeds.
Kweon et al. (18) estimated the total safety effects of speed limit changes on high-speed
roadways by using traffic detector data and Highway Safety Information System data from 1993
to 1996. The study used a sequential modeling approach in which average speed and speed
11
variance models were first estimated based on the design, use and speed limit information; crash
counts were estimated based on the speed estimates, design, and use variables. 63,937
homogeneous highway segments along Washington State’s 7 Interstates and 143 state highways
provided the data for 4 years. Results indicated lower nonfatal crash rates up to 55 mph speed
limit. On the other hand, fatality rates were unresponsive to speed limit changes.
2.3 PCE RELATED STUDIES
The effect of truck lane restriction and differential speed limit policies can also be measured
by estimating the passenger car equivalents of trucks, ET. The term “Passenger car equivalent”
was first introduced in HCM 1965 to define the effect of trucks and buses in the traffic stream. It
was defined as “the number of passenger cars displaced in the traffic flow by a truck or a bus,
under the prevailing roadway and traffic conditions”. HCM 1950 used a single factor of 2.0 to
account for the impact of heavy vehicles on multi-lane highways. However the most recent
definition of PCE is in HCM 2000 and which is defined as “the number of passenger cars that are
displaced by a single heavy vehicle of a particular type under prevailing roadway, traffic and
control conditions”. The literature shows a few studies that used to capture the effect of
trucks on traffic operation during similarly unusual operating conditions. For example, Chitturi
and Benekohal (19) evaluated the impact of work zones on and found that values
decreased as the truck percentage increased and also increased as the traffic volume increased.
Another study by Ahmed et al. (20) concluded that the effect of heavy vehicles on traffic is
more noticeable during congestion than during under-saturated conditions. This study attempts
to determine how the restriction policies affect the impact of trucks on traffic operation. The
effect is primarily captured by estimating values for a four-lane rural freeway segment under
different traffic conditions and compliance rates.
Nicholas (22) et al identified the impact of truck lane restriction strategies where gradient is
also applicable. This study evaluated the impact of these restrictions on the traffic operation and
safety for several combinations of traffic and geometric characteristics using the simulation
program PARAMICS that simulates the different traffic and geometric characteristics. The
12
measurement of effectiveness (MOE) for safety employed in this study is conflict, a surrogate
measurement for traffic crash. The impact of different lane restrictions in terms of the above
MOE was evaluated for different traffic conditions (volume, truck percentage) and geometric
characteristics (gradient, speed limit, intersection density). ANOVA analysis of the simulation
results indicates that truck lane restrictions show significant impacts upon all kinds of conflicts
related to trucks.
A study conducted by Werner B et al (23) found that capacity is not limited by the length of
a gradient. The limiting impact of the gradient on capacity usually reaches its maximum at a
gradient length of around 500 m. Travel velocity, however, is significantly influenced by the
degree of gradient and the length of the upgrade (up to L ≤ 4000 m) as well as by the proportion
of trucks.
Another study conducted by Lily et al (24) observed that for freeway sections, PCEs remain
mostly unchanged or even decrease (by 1 unit) with increasing traffic flow, especially the low
performance trucks. For arterials and two-lane highways, there is no discernible trend between
PCEs and level of traffic flow. For freeway sections, PCEs remain mostly unchanged, and
sometimes they increase (by 1 to 5 units) with an increasing percentage of trucks in the traffic
stream. This occurs mostly for long and steep grades. For arterials, however, PCEs mostly
decrease with an increasing percentage of trucks. For arterials, grade was not considered,
because NETSIM (Network Simulation) does not account for the effects of grade. For two-lane
highways, there is no visible trend between the percentage of trucks and PCE values. Generally
major differences in PCEs occur for the longer and steeper grades. This is because when grades
are considered the trend has shown irregularities which are not suitable for this study.
From the past research on speed limit restrictions, it can be concluded that this alternative
may reduce crashes involving trucks rear-ending other vehicles and others have concluded that
differential speed limits increases speed variances and lane changes on roadways, which may in
turn increase the chance of a crash occurring. The actual benefit of implementing this strategy is
still questionable even though it has been proved to be beneficial at some locations. Some other
13
studies were conducted when grades play an important role in traffic operational efficiency.
2.4 SUMMARY
Contradictory conclusions were found on the safety benefits and effectiveness for both
truck lane restrictions and DSL in literature. Studies on truck lane restrictions and DSL were
conducted on flat and upgrade segments, but not on elevated segments such as the Atchafalaya
segment. Previous findings show that truck lane restrictions increases speed differentials and
reduce density on steep upgrades. Some studies found that truck lane restrictions do have a
positive impact on freeway safety. No study has been conducted on the effectiveness of
implementing both trucks-lane restriction and DSL on segments with two lanes in each
direction as Atchafalaya segment of I-10.
The Highway capacity manual could only provide estimates for freeway capacity that are
calibrated to a set of “ideal” conditions. But most of the time the prevailing conditions are not
ideal as the traffic stream contains a mixture of passenger cars as well as trucks and RV’s. So
HCM 2000 utilizes passenger car equivalents to account for the presence of heavy vehicles in
the traffic stream. Using the PCE’s, a non-homogenous mix of vehicles can be expressed in a
standardized unit of traffic. But these PCE’s were not quantified when the restriction policies
are in place when published in HCM and so it is a question still unanswered and hence the case
for this study.
14
3. STUDY SECTION AND DATA COLLECTION
3.1 STUDY SECTION
The study was conducted on an 18-mile four-lane elevated rural segment of I-10 over the
Atchafalaya basin between the city of Baton Rouge and Lafayette in Louisiana as shown in
Figure 3. The sites where the data were collected are shown in the figure. The 30-second traffic
counts, spot speeds, vehicle classification, and occupancy data was used to estimate the
distribution of speeds under free-flow conditions, by vehicle classification, under the existing
operational restrictions, trucks restricted to travel on the right lane at 55 mph and cars allowed to
travel on both lanes at 60 mph.
Figure 3: Study section where data collected are mentioned as Site #
3.2 DATA COLLECTION
The segment under investigation was monitored for 4 months from June 11, 2007 to Sept 22
, 2007 to capture the changes of traffic characteristics over time in terms of flow rates, speed
distributions, and vehicle composition. A true presence detector, Remote Traffic Microwave
Sensor (RTMS), was used to collect traffic data. Since the traffic characteristics vary by lane
and direction of travel, four sites on both directions were selected. At each site, the RTMS was
attached to the poles of the speed limit sign to observe the traffic of both lane and collect data
every 30 seconds. The data however has some voids which have been neglected as they are very
minimal and dispersed and would not affect the overall data.
15
3.2.1 DATA COLLECTION EQUIPMENT - RTMS
The RTMS radar used for the data collection is a low-cost, general-purpose, all-weather
traffic sensor that is currently used on segment of I-10 and I-12 in Louisiana. Four sensors were
installed on the selected four observation sites, and detected the presence of traffic in both lanes.
As a true presence traffic detector, RTMS devices provide automobile average speed,
vehicle counts, occupancy, and additional classifications for long vehicle counts in the detection
zones. The length-based vehicle classifications setup for RTMS in the study was shown in Table
1. The data were collected every 30 seconds over the monitoring period.
The RTMS devices collected data and sent the information wirelessly to the antenna
mounted at the Butte La Rose tower. The data were stored in the cluster controller and
transferred to the PC in the office from the cellular modem in the tower. The transferred data was
stored in MS Access format.
Table 1: Vehicle classifications
Vehicle classification Length
Regular = <26'
Midsize = >26' - <36'
Long = >36' - <56'
Extra-Long = >56' - <76'
3.2.2 DATA COLLECTED
The description of the data collected by RTMS is shown in Table 2. The speed measured by
RTMS is the average speed of all vehicles passing by during the 30 seconds time interval. As for
the four vehicle categories, the Long and Extra_Long vehicle counts are considered as truck
volume in the following data analysis.
With the six month data available which amounted to a total of 1 million data, this was used
for analysis in the study, since it is long enough to capture the change of traffic characteristics
over time and one million records provide enough samples to perform statistical analysis. The
raw data transferred from the RTMS devices between June 11 and September 26, 2007 has
16
1,647,914 records. After deleting the void measurements, 1,298,504 records were left. The
variables used in the analysis include Time, Zone, Lane, total volume, volume of different
vehicle types, traffic speed, and Occupancy. Among the variables analyzed, Volume and Speed
are numerical, while Zone, Lane and Occupancy are categorical.
Table 2: Description of data collected by RTMS Column Name Description
DateTimeStamp The Server data time of the data insertion.
RTMS_NETEORK_ID TRMS Network ID of the reporting TRMS unit.
RTMS_NAME
The name associated with the RTMS as defined by
Station Manager Network Configuration
Lane RTMS zone number for which the statistics are reported
Speed Average speed of the last reported message period.
Volume Volume counts for the last reported message period.
Vol_Mid
Midsize vehicle volume counts for the last reported
message period.
Vol_Long
Long vehicle volume counts for the last reported message
period.
Vol_Extra_Long
Extra long vehicle volume counts for the last reported
message period.
Occupancy Lane occupancy for the last reported message period.
Speed unit 0/false if km/hr, 1/ture if mph.
17
4. METHODOLOGY
4.1 INTRODUCTION
The data collected from the Atchafalaya test section is utilized for this research study to
derive the PCE’s under a range of restricting conditions and composition of the traffic mix. To
accomplish this, the freeway is modeled using the simulation tool, VISSIM. The input variables
are explicitly defined as these variables form the backbone of the whole study and also control
the whole simulation environment to get the desired output. After the control variables are
specified, the different case scenarios are introduced. The different case scenarios are a
combination of restriction policies in place for trucks and the compliance rate of the vehicles to
these policies. A base case scenario is also defined as a “ideal” condition where the freeway is
considered having only passenger cars and this base case is used to compare all the other
scenarios defined. These scenarios are run in the simulation tool and the output is quantified in
terms of PCE’s. These output values are compared against the HCM 2000 PCE values.
4.2 CODING AND INPUT FOR SIMULATION
To evaluate the impact of operational restrictions, the study segment was coded in
VISSIM. The whole east-bound two-lane direction of traffic was modeled in VISSIM, including
the two off-ramps located in the middle of the restricted section. The following section
describes the methodology developed to evaluate the performance of the modeled freeway
segment under various levels of operational restrictions. While modeling this freeway segment
several assumptions were considered. Assumptions were that the typical amount of traffic
exiting on the middle ramps is low, and moreover, since only the basic freeway segment case is
considered in this study it was assumed that all the traffic entering the 18-mile segment goes
through and no vehicle exits in the middle of the section. In addition, to represent freeway
conditions more realistically and independent of the simulator characteristics an upstream two-
mile long three-lane extension was coded before the actual test section begins. The next step
involves defining the input variables which controls the simulation process.
18
4.3 INPUT/ CONTROL VARIABLES
Several control variables were defined for this study to effectively study the sensitivity of
the analysis done using this tool. A couple of the variables i.e. speed distributions and vehicle
types formed the input parameters for the simulation tool while the other variables were used
specifically for this study.
4.3.1 SPEED DISTRIBUTIONS
The distributions of the desired free-flow speeds for passenger cars and trucks were
derived from I-10 data. The data collected from the restricted section of I-10 over the
Atchafalaya Swamp was used to estimate the free-flow speed for each vehicle type at flow rates
less than 1300 pcphpl(passenger car per hour per lane), as specified by the HCM 2000. A
passenger car equivalent factor for trucks was assumed to be 1.5 since the elevated section does
not have any significant grades, neither length nor magnitude wise. These distributions are
marked as base field data in Figure 4 and Figure 5. It is interesting to note that even though the
posted speed limit for cars and trucks are 60 mph and 55 mph, respectively, the observed mean
speed values for these distributions are 64.9 mph and 59.4 mph, respectively. Basically, for flow
rates less than 1300 pcphpl the mean values are nearly 5 mph greater than the posted speed
limits. In addition, to account for a more relaxed speed enforcement scenario two more
distributions were derived from the observed one by assuming 10% and 20% respectively,
increased in the free-flow speed for all vehicles. These distributions were named as base + 10%
and base+20% respectively. These calculated distributions lead to mean speed values of nearly
71 mph and 78 mph for cars and 66 mph and 72 mph for trucks.
The free flow speed(ffs) for both passenger cars and the trucks were also taken into
account and marked as ffs – 70 mph, which is an ideal condition to be considered.
19
Figure 4- Speed Distribution curves for Passenger cars from field data
Figure 5- Speed Distribution curves for Trucks obtained from field data
4.3.2 VEHICLE TYPES
The vehicle population in VISSIM is categorized into vehicles types. A single type gathers
vehicles that share common vehicle performance attributes. These attributes include model,
minimum and maximum acceleration, minimum and maximum deceleration, weight, power, and
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
40 45 50 55 60 65 70 75 80 85 90 95 100 105
base field data
base +10%
base+20%
ff-70mph
Per
centa
ge
of
pas
senger
car
s
Speed in Mph
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
40 45 50 55 60 65 70 75 80 85 90 95 100
base field data
base +10%
base+20%
ff-70mph
Per
cen
tag
e o
f T
ruck
s
Speed in Mph
20
length. All of these, except for model and length, are defined in VISSIM with probabilistic
distributions (as opposed to scalars) as seen in Figure 6. In order to account for various
compliance rates to truck lane restriction in VISSIM trucks were coded as Heavy Goods Vehicle
classes, namely HGV-left and HGV-right, to account for the trucks on the left and right lanes.
The freeway segment was coded such that only trucks belonging to HGV-left are prohibited to
travel onto the left lane, while the other vehicles can freely chose their travel lane. The vehicle
specification for Car and truck type is identical to that of the default CAR and TRUCK type in
the basic VISSIM Model.
Figure 6: Speed distribution curve in VISSIM, from field data
Figure 7: Basic Vehicle types used for this research
The other default vehicle types in VISSIM are as shown in the above table. For this study the
vehicle types were custom-defined into the software to account for the Truck type and Car type
to replicate the actual conditions. A vehicle class represents a logical container for one or more
21
previously defined vehicle types. A vehicle type can also be part of several vehicle classes, thus
“overlapping” classes are possible. Thus the types represent a clear picture.
4.3.3 PERCENTAGE OF TRUCKS IN THE TRAFFIC MIX
Despite their smaller proportion of vehicular traffic, heavy vehicles are known for their
important impacts on traffic flow. Two factors account for these impacts. First, the dimensions of
trucks are generally larger than those of passenger cars and therefore the average space taken up
by a truck is greater than that taken up by a typical passenger car. Second, the operational
characteristics (acceleration, deceleration, maneuverability, etc.) of these heavy vehicles are
different from those of passenger cars. As a result, heavy vehicles are believed to have a physical
effect on nearby vehicles and a psychological impact on the drivers of those vehicles. Hence
varying the percentage of the trucks in the mix, the different scenarios are formulated to study its
impact on the PCE values. For this study truck composition ranging from 10% to 40% is used.
4.3.4 RATE OF COMPLIANCE TO THE RESTRICTION POLICIES
Compliance with the restrictions was broadly classified into three, i.e. when 100% of the
vehicles comply with the restrictions and similarly when 75% and when only 50% of the
vehicles complies with the restriction policies. The proportion of vehicles traveling in the
restricted lane was examined to determine the degree of non-compliance with the restrictions.
The number and percentage of trucks impeding flow were also examined to determine the extent
to which slow-moving trucks were reducing the speed of other traffic.
4.4 SIMULATION CASE SCENARIOS
Each scenario represents a unique combination of desired speed distribution, traffic
composition, compliance rate to truck lane restriction, compliance rate to differential speed limit
and different combination of upgrades. To account for stochastic variations in the model each
scenario was simulated ten times with different random seeds. The Base case scenario is taken
with the total flow input into the freeway segment assumed to be 1300 pcphpl, which is close to
the estimated throughput of a two-lane freeway and to account for free-flow conditions. All
simulated scenarios are described in the following table. The scenarios are explained in Table
22
5.Table 5The speed distributions A, B, C and D are from Table 3. The characteristics of the
observed distributions are indentified in Table 3 as Speed Distribution A. In addition, to
simulate various compliance rates to the same differential speed limit, two more speed
distributions were computed. Speed Distribution B and C were derived to provide a higher and
a lower compliance rate, respectively. The last row in Table 3 shows Speed Distribution D that
assumes a speed limit of 70 mph common to all vehicles. Speed Distribution D is used to
simulate no speed restriction conditions. All the above 12 scenarios with 4 types of speed
distributions namely A,B, C and D as explained in Table 5 with the gradient combinations brings
up the total number of scenarios to 432, excluding one base case scenario which considers free
flow conditions with no trucks. Speed distribution A corresponds to base speed obtained from
field data, B represents base speed+10%, C represents base speed + 20% and D is the free flow
speed for all the scenarios respectively.
Table 3: Distribution of desired speeds for passenger cars (PC) and trucks
Speed Distribution
Compliance to
differential speed
limit [%]
Mean Speed
[mph]
Median Speed
[mph]
85th Percentile
[mph]
Speed Distribution A
(observed compliance)
PC 35 65 63 74
Truck 25 60 58 64
Speed Distribution B
(high compliance)
PC 50 60 60 67
Truck 50 55 55 61
Speed Distribution C
(low compliance)
PC 15 70 74 77
Truck 15 65 55 70
Speed Distribution D
(no restriction)
All
veh. N/A 70 70 80
Table 4: Grades and their respective lengths for each scenario
Upgrade <2% (1%) >3-4%(3%) <5-6%(5%)
Length 0.25 0.5 1 0.25 0.5 1 0.25 0.5 1
Each of the 48 scenarios has 9 different combinations of gradients and their lengths. So the total
number of scenarios considered is 432. The grades and their lengths are shown in Table 4. The 3
different types of upgrades and their corresponding lengths are selected to be consistent with
23
HCM. The location of the grades were selected near the exit and entry and in the middle of the
section
Table 5: Case Scenarios for each speed distribution
Scenario Speed limit of
trucks/cars [mph] Compliance to truck
lane restriction [%] Percentage of Trucks [%]
1 55/60 100 10
2 55/60 100 20
3 55/60 100 30
4 55/60 100 40
5 55/60 75 10
6 55/60 75 20
7 55/60 75 30
8 55/60 75 40
9 55/60 50 10
10 55/60 50 20
11 55/60 50 30
12 55/60 50 40
4.5 ESTIMATING ET USING SIMULATION
After the scenarios are defined, the input variables are introduced into VISSIM to create the
coding for the required simulations.
4.5.1 FREEWAY CODING IN VISSIM
This freeway section was built into VISSIM using base maps in bit map format. The study
section is built in VISSIM through a series of links and connectors. Links are generally straight
or follow the curvature of the road. Connectors, which are used to connect links, are typically
used to model turning areas and lane expansions and contractions. In VISSIM, the creation of the
freeway section is fairly simple through the use of a graphical interface and an aerial photograph
in the background. For this study the map was taken from MapQuest and Google maps.
4.5.2 CALIBRATION AND VERIFICATION
Since simulations are mathematical simplifications of real-world phenomena, their
capability of replication must be verified prior to application to the real world. Since the default
model parameters in VISSIM were not calibrated based on those in United States, this model has
to be calibrated for this research study and validated. Calibration is the process by which the
24
individual components of the simulation model are refined and adjusted so that the simulation
model accurately represents field measured or observed traffic conditions. VISSIM simulation
model contains default values for each variable which permits a range of user-applied values for
each variable. In some cases, the variables affect the entire network while others are specific to
individual roadway segments or nodes. Changes to these variables during calibration should be
based on field-measured or observed conditions. In other words, a change in the variables
should be justified and defensible. Unfortunately, the user manual for VISSIM provides little or
no information about the source or appropriateness of the default parameters, nor does it provide
substantial guidance on how the user should modify these parameters for different types of
conditions. Therefore, the user has a greater responsibility for ensuring that appropriate changes
are made that are based on field-measured data and not exclusively on engineering judgment.
Any microscopic model must be verified after network coding and before proceeding
further. Verification involves visual examination of coded network to ensure that the coded
network represents actual conditions.
For this specific research, the simulation is calibrated by taking into the consideration the
ideal scenario when all the vehicles are passenger cars and so the ET should also be ideally 1.0
corresponding to the “ideal” traffic mix. VISSIM was allowed to run for 10 different random
seeds with a traffic input of only passenger cars and the ET was derived from the simulation
results. The ensuing results did not vary much from the expected Et of 1.0. The values ranged
from 0.98 to 1.02 and hence the model was considered suitably calibrated for this study.
4.5.3 ESTIMATION OF ET FROM SIMULATION OUTPUT
To estimate the value of , two main assumptions were made in the simulation process as
following;
1) The freeway segment was considered homogeneous such that all vehicles entering the
segment exit at the end of the segment (i.e. no traffic exiting/entering in the middle of the
segment).
25
2) The values of were assumed to be independent of the traffic flow rates.
Assuming input flow as 1300 pcphpl to maintain free-flow conditions, the simulation is
run for a period of 1 hour and the time taken T, for the beginning of the simulation where the
first vehicle enters to the time the last vehicle exits the section is determined, which would be
subsequently used in all the other case scenarios. The base case scenario was defined to account
for ideal traffic conditions that assume a traffic composition with no trucks and a speed limit of
70 mph.
For all the other scenarios, the total number of vehicles that arrive at the end of the section
during the time T is observed and named as NTi and N PC
i. where represents the total
throughput in the base case with no trucks (NT0). represents the total number of passenger
cars in the corresponding scenario and is the total number of trucks. For the base case
scenario the NT0 is zero and NPC
0 is the throughput multiplied by the number of lanes.
Given , the set of simulation scenarios from 1 to 48, for each scenario , the number of
vehicles passing a point at the end of the section, is measured over a 3-mile long section.
Similarly, represents the 15-minute flow rate of the Base Case scenario. All flow rates can be
converted to pcphpl if truck flows are converted to passenger car units through . This
computation method to estimate is shown in equations (1)-(3).
(1)
– = (2)
– / (3)
26
5. SIMULATION RESULTS
This chapter summarizes all the results from the estimation of ET using the methodology
explained in the preceding chapter. The output results were broadly analyzed into two based on
whether upgrades were introduced into the section and when it is not.
5.1 WHEN NO GRADIENTS ARE TAKEN INTO CONSIDERATION
This section explains the observations made from the simulation results when restriction
policies are in place without any gradients in the section. All the simulation scenarios are
explained in the table below.
Table 6: Simulation scenarios
Sce
nar
io
Sp
eed
dis
trib
uti
on
Sp
eed
li
mit
of
tru
cks/
cars
[mp
h]
Co
mp
lian
ce
to t
ruck
lan
e
rest
rict
ion
[%]
Per
cen
tag
e
of
Tru
cks
[%]
Sce
nar
io
Sp
eed
dis
trib
uti
on
Sp
eed
li
mit
of
tru
cks/
cars
[mp
h]
Co
mp
lian
ce
to t
ruck
lan
e
rest
rict
ion
[%]
Per
cen
tag
e
of
Tru
cks
[%]
1 A 55/60 100 10 25 C 55/60 100 10
2 A 55/60 100 20 26 C 55/60 100 20
3 A 55/60 100 30 27 C 55/60 100 30
4 A 55/60 100 40 28 C 55/60 100 40
5 A 55/60 75 10 29 C 55/60 75 10
6 A 55/60 75 20 30 C 55/60 75 20
7 A 55/60 75 30 31 C 55/60 75 30
8 A 55/60 75 40 32 C 55/60 75 40
9 A 55/60 50 10 33 C 55/60 50 10
10 A 55/60 50 20 34 C 55/60 50 20
11 A 55/60 50 30 35 C 55/60 50 30
12 A 55/60 50 40 36 C 55/60 50 40
13 B 55/60 100 10 37 D 70 100 10
14 B 55/60 100 20 38 D 70 100 20
15 B 55/60 100 30 39 D 70 100 30
16 B 55/60 100 40 40 D 70 100 40
17 B 55/60 75 10 41 D 70 75 10
18 B 55/60 75 20 42 D 70 75 20
19 B 55/60 75 30 43 D 70 75 30
20 B 55/60 75 40 44 D 70 75 40
21 B 55/60 50 10 45 D 70 50 10
22 B 55/60 50 20 46 D 70 50 20
23 B 55/60 50 30 47 D 70 50 30
24 B 55/60 50 40 48 D 70 50 40
27
5.1.1 TRUCK LANE RESTRICTION ONLY
Twelve simulation scenarios, scenario 37 to 48 were used to estimate the effect of
compliance rate to the truck lane restriction. Figure 8 depicts a general trend of decrease in
with the increase in the percentage of trucks. For example, if 100% compliance to truck lane
restriction is assumed, is changing between 1.46 and 1.80 while varies between 10% and
40%.
A two sample T-test was conducted to investigate if the percentage of trucks leads to
statistically significant differences of the mean values between two samples of the results, A
and B. The sampling for this statistical analysis is explained in the Table 8. PT represents the
percentage of trucks in the mix. The null hypothesis states that ETA > ET
B where ET
A and ET
B
represents the mean value for for sample A and B, where A and B represents the ET obtained
from considering the desired Speed distributions A and B as explained Table 3. As can be seen
from Table 8, regardless of the compliance to truck lane restriction, with a few exceptions,
shows statistically different values for different values. Since each simulation scenario was
performed ten times with different random seeds, the critical value of T-stat from the two-sample
T-test was calculated for eighteen degrees of freedom at 0.05 level of confidence as,
.
Table 7 also shows that the higher the compliance to the lane restriction the larger the
difference between values. But some values were observed not statistically different. Case
A, B and C represents the compliance rates of 100%,75% and 50% respectively. Also using the
mean values of a linear approximation was derived to estimate for intermediate traffic
compositions and R- Square value was determined, which was very close to 1. For each of the
three compliance rates to truck lane restriction a fitted equation is listed in. However, caution
must be used for values exceeding 40%. So regardless of the compliance rates, the value of
ET is on the decline as the percentage of trucks increases.
28
Later all the case scenarios are compared with the base case scenario and the trend is
studied. The analysis shows that the downward trend may be caused by the platooning effect by
the trucks when combined with the speed restriction and also the truck lane restriction worsens
the interaction between other vehicles causing the efficiency to come down considerably.
Figure 8: Effect of % trucks on under various compliance rates to truck lane
restriction.
Table 7: T-test for significant difference of under truck lane restriction only
1
1.2
1.4
1.6
1.8
2
10 20 30 40
100-0
75-25
50-50
ET
Percentage of trucks
Compliance to Truck
Lane Restriction [%]
(Sample A)
[%]
(Sample B)
[%]
(Sample A)
(Sample B)
T-statistic*
100% 10% 20% 1.82 1.65 5.30*
100% 10% 30% 1.82 1.52 9.20*
100% 10% 40% 1.82 1.27 9.84*
100% 20% 30% 1.65 1.52 3.26*
100% 20% 40% 1.65 1.27 6.25*
100% 30% 40% 1.52 1.27 4.07*
75% 10% 20% 1.67 1.64 0.33
75% 10% 30% 1.67 1.54 2.47*
75% 10% 40% 1.67 1.32 3.74*
75% 20% 30% 1.64 1.54 1.29
75% 20% 40% 1.64 1.32 2.87*
75% 30% 40% 1.54 1.32 2.23*
50% 10% 20% 1.67 1.63 0.82
50% 10% 30% 1.67 1.55 2.02*
50% 10% 40% 1.67 1.43 3.03*
50% 20% 30% 1.63 1.55 1.23
50% 20% 40% 1.63 1.43 2.37*
50% 30% 40% 1.55 1.43 1.29 *significance level of 0.05
29
Table 8: Approximation of under truck lane compliance rates
Compliance to Truck Lane
Restriction [%] Fitted Equation R-Square
100 = 1.9593 0.1425 * 0.9933
75 0.8771
50 = 1.7794 0.1013 0.9109
5.1.2 COMBINED EFFECT OF BOTH RESTRICTION POLICIES
The combined effect of compliance rate to lane restriction for trucks and to differential
speed limits was analyzed based on the remaining 36 simulation scenarios from 1 to 36.
Similarly, it was found that is impacted by changes in the truck composition in the traffic mix
(i.e. ). The values of are depicted in Figure 9, Figure 10 and Figure 11. The trend line in
these figures show that decreases with , regardless of the compliance rates to truck lane
restriction and to differential speed limit restrictions. As expected, for a given traffic
composition value, ET seems to decrease with the increase in the compliance to differential speed
limit and to track lane restriction regardless of the combination of the policies in effect. This can
be explained by the fact that the combination of the two policies lead to reduced interaction
between passenger cars and trucks. Consequently, reduced vehicle interactions lead to smaller
impact on the value of PCE for trucks.
A two sample T-test was conducted to investigate if the percentage of trucks leads to
statistically significant differences of the mean values between two samples of the results, A
and B. PT represents the percentage of trucks in the mix. The null hypothesis states that ETA >
ETB where ET
A and ET
B represents the mean value for for sample A which has desired speed
distribution A and sample B which has desired speed distribution B, respectively. Case A, B and
C represents the compliance rates of 100%, 75% and 50% respectively. Table 9 shows the
results of the T-test applied for all the simulation scenarios based on the desired Speed
Distribution A (i.e. the speed distribution observed on the study section). It can be seen that if
the compliance to lane restriction is 100%, for all traffic compositions has statistically
30
significant different values to significance level of 0.05. The significance may drop if the
percentage of trucks are high (e.g. changes from 30% to 40%) as not all drivers comply with the
truck lane restriction. Nevertheless, for the cases with lower percentage of trucks (10% and
20%) yields significantly different values regardless of the compliance to the truck lane.
Figure 9: Percentage of trucks plotted against ET for 100% compliance
Figure 10: Percentage of trucks plotted against ET for 75% compliance
Figure 11: Percentage of trucks plotted against ET for 50% compliance
0.000
0.500
1.000
1.500
2.000
10 20 30 40
case a
case b
case c
base case
0.000
0.500
1.000
1.500
2.000
10 20 30 40
case acase bcase cBase case
0
0.5
1
1.5
2
10 20 30 40
case a
case b
case c
Base Case
31
Table 9: T-test for ET under truck lane restriction and differential speed limit
Compliance to
Truck Lane
Restriction [%]
(Sample A)
[%]
(Sample B)
[%]
(Sample A)
(Sample B)
T-statistic*
100% 10% 20% 1.80 1.66 3.26*
100% 10% 30% 1.80 1.58 5.25*
100% 10% 40% 1.80 1.46 3.69*
100% 20% 30% 1.66 1.58 1.97*
100% 20% 40% 1.66 1.58 2.17*
100% 30% 40% 1.58 1.58 1.23
75% 10% 20% 1.63 1.63 0.00
75% 10% 30% 1.63 1.52 2.63*
75% 10% 40% 1.63 1.41 2.81*
75% 20% 30% 1.63 1.52 2.63*
75% 20% 40% 1.63 1.41 2.81*
75% 30% 40% 1.52 1.41 1.33
50% 10% 20% 1.43 1.61 -2.09*
50% 10% 30% 1.43 1.52 -0.96
50% 10% 40% 1.43 1.40 0.34
50% 20% 30% 1.61 1.52 1.34
50% 20% 40% 1.61 1.40 2.90*
50% 30% 40% 1.52 1.40 1.53 *significance level of 0.05
Table 10: Approximation of ET for different compliance rates to differential speed limits
Speed Distribution
Compliance to
Truck Lane
Restriction [%]
Fitted Equation
( where X - ,y - R- Square
A
(observed compliance)
100 y = -0.1782x + 2.0099 0.9837
75 0.8
50* y = -0.0499x2
+ 0.1114x +
1.5821 0.9866
B
(high compliance)
100 y = -0.1252x + 1.9647 0.9559
75 y = -0.1138x + 1.8276 0.88
50* y = -0.0758x2
+ 0.3598x +
1.1591 0.8809
C
(low compliance)
100 y = -0.1555x + 1.9579 0.9338
75 y = -0.1218x + 1.8427 0.7986
50* y = -0.0388x2
+ 0.0815x +
1.6002 0.9968
* Polynomial fitting equations were used
32
5.2 EFFECT OF RESTRICTIONS WHEN GRADES ARE APPLIED
The combined effect of compliance rate to truck lane restriction and to differential speed
limit was again analyzed for sections with grades for all the scenarios from 1 to 36 with different
combinations of gradient lengths. The lengths of grades were selected in line with HCM as 0.25,
0.5 and 1 miles respectively. Similarly, it was found that is impacted by changes in the traffic
composition when restriction policies are applied on grades. For a given traffic composition
value ET seems to decrease with PT, regardless of the compliance rates to truck lane restriction
and to differential speed limit restriction with gradients also in consideration.
A one sample T-test was conducted to investigate if the ET mean values leads to
statistically significant differences of the as given in HCM 2000. The null hypothesis states
that ETSIMU
> ETHCM
ETSIMU
and ETHCM
represent the mean value for for simulated scenarios
and the corresponding HCM values, respectively. Table 11 shows the results of the T-test applied
for all the simulation scenarios based on the desired Speed Distribution A (i.e. the speed
distribution observed on the study section). It can be seen that if the difference in values
increases as the grade increase with compliance to all lane restrictions
For all traffic compositions ET has statistically significant different values 0.05 level of
significance. The significance may drop if the grades are less (e.g. from 0.25 to 0.5 Mile) with
lower truck percentages as not all drivers comply to the truck lane restriction either. The shaded
values in Table 11 shows which are significantly different from the corresponding values in
HCM 2000 for the truck composition and length of grade. The algebraic difference of the values
are taken into account to formulate the following table.
The results show that values differ for higher truck percentages which can be interpreted as
higher truck percentages on grades nullified the effect of speed limit differentials as the truck
speed is lower than posted speed limit and hence the interaction between passenger cars and
trucks increases.
33
Table 11: Significance on upgrades compared to HCM
50-50 Truck
Compliance
55-60 Speed
restriction
Percentage of trucks ET
Upgrade Length 10 20 30 40
<2 (1%)
0.25 0.05 0.67 0.01 0.00
0.5 0.20 0.36 0.01 0.00
1 0.09 0.22 0.01 0.00
>3-4(3%)
0.25 0.76 0.00 0.00 0.00
0.5 0.00 0.36 0.01 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%)
0.25 0.00 0.36 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.01 0.00 0.00 0.00
Upgrade Length 10 20 30 40
<2 (1%)
0.25 0.29 0.29 0.01 0.00
0.5 0.40 0.40 0.01 0.00
1 0.97 0.97 0.01 0.00
>3-4(3%)
0.25 0.07 0.07 0.02 0.00
0.5 0.00 0.40 0.01 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%)
0.25 0.00 0.00 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.01 0.01 0.00 0.00
Upgrade Length 10 20 30 40
<2 (1%)
0.25 0.10 0.58 0.02 0.00
0.5 0.05 0.73 0.03 0.00
1 0.07 0.41 0.02 0.00
>3-4(3%)
0.25 0.85 0.01 0.00 0.00
0.5 0.00 0.73 0.03 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%)
0.25 0.00 0.00 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.00 0.00 0.00 0.00
Upgrade Length 10 20 30 30
<2 (1%)
0.25 0.10 0.58 0.02 0.00
0.5 0.05 0.73 0.03 0.00
1 0.07 0.41 0.02 0.00
>3-4(3%)
0.25 0.85 0.01 0.00 0.00
0.5 0.00 0.73 0.03 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%)
0.25 0.00 0.00 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.00 0.00 0.00 0.00
Upgrade Length 10 20 30 40
<2 (1%)
0.25 0.10 0.61 0.00 0.00
0.5 0.27 0.30 0.00 0.00
1 0.35 0.49 0.00 0.00
>3-4(3%)
0.25 0.07 0.17 0.00 0.00
0.5 0.45 0.30 0.00 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%)
0.25 0.00 0.00 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.00 0.00 0.00 0.00
Upgrade Length 10 20 30 30
<2 (1%) 0.25 0.37 0.61 0.00 0.00
0.5 0.19 0.30 0.00 0.00
1 0.09 0.49 0.00 0.00
>3-4(3%) 0.25 0.07 0.17 0.00 0.00
0.5 0.00 0.30 0.00 0.00
1 0.00 0.00 0.00 0.00
<5-6 (5%) 0.25 0.00 0.00 0.00 0.00
0.5 0.00 0.00 0.00 0.00
1 0.00 0.00 0.00 0.00
34
6. CONCLUSIONS
6.1 STUDY CONCLUSIONS AND FUTURE WORK
This research study presented a methodology to quantify the effect of truck lane restriction
and differential speed limit policies on the passenger car equivalent for trucks, . The study is
preformed primarily for an elevated level terrain basic freeway segment. A sensitivity analysis
was conducted using microscopic simulation and real-world traffic data collected from a 18-mile
freeway segment of I-10 in the State of Louisiana. The study section is a four-lane freeway that
implements both policies, truck lane restriction (i.e. trucks are allowed to travel on the right lane
only) and differential speed limit (i.e. cars and trucks speed limit are 60 mph and 55 mph,
respectively). Data collected from the study section reveals relatively moderate compliance to
the implemented policies, and simulation is used to derive values under various compliance
rates to the two restrictions.
It was found that if only truck lane restriction is used, ET has statistically significant different
values for changing percentage of trucks in the traffic composition. The value of ET ranges
between nearly 1.4 and 1.8. Also, ET shows a linear decrease with the percentage of trucks.
Linear models were derived to calculate for three compliance rates to truck lane restriction:
100%, 75%, and 50%, respectively. Similarly, a linear relationship between ET and the
percentage of trucks, can be defined if both policies are in place (i.e. truck lane restriction
and differential speed limit). The values of ET for different traffic composition values and
various compliance rates to the two restrictions range between 1.3 and 1.8, and are statistically
different. In addition, linear approximations were derived to estimate based on . Even
though, the fitted linear models were very good, caution should be used in applying for values
greater than 40%.
Although the estimated values of are not much different from 1.5, which is the value
currently suggested by HCM 2000, they are statistically significantly different. Hence, its impact
on estimating the level of service for the basic freeway sections that operate under these non-
35
typical conditions can be observed especially for cases at the borderline between two adjacent
levels of service. A more complete analysis is thought to be conducted in order to test for similar
operational policies implemented on freeways segments with significant grades. Also, future
work should test for passenger car equivalents of trucks and other vehicle types as well (e.g.
RVs) on freeway segments with six lanes or more and various differential speed limit values and
trucks lane restrictions (e.g. trucks restricted to travel on two lanes vs. one lane).
36
REFERENCES
1 Highway Capacity Manual. TRB, National Research Council, Washington, D.C., 2000.
2 Zavoina, M.C. Operational Evaluation of Truck Restrictions on I-20 in Texas,
Transportation Research Record 1320, TRB, National Research council, Washington,
D.C., pp 24-31, 1991
3 Hoel, L. A. and J.L. Peek. A Simulation Analysis of Traffic Flow Elements for
Restricted Truck Lanes on Interstate Highways in Virginia, VTRC-99-R1, Final report,
Virginia Transportation Research Council, Charlottesville, VA, 1999.
4 Mussa, R. and Price, G. Quantify the Effects of Raising the Minimum Speed on Rural
Freeways and the Effects of Restricting the Truck Lanes Only in the Daytimes, BC-352,
Research Final report, Vol.2, Traffic Operation Office, Department of Transportation,
Tallahassee, FL, 2004.
5 Borchardt, D. TTI Evaluates Lane Restrictions for Houston Demonstration Project. Texas
Transportation Researcher, Vol. 38, No.1, 2002, pp. 9-10, 2002.
6 Zeitz, R. Low Cost Solutions Yield Big Savings, Public Roads, Vol 67, No. 3, 2003.
7 Gan. A. and S. Jo. Operational Performance Models for Freeway Truck-Lane
Restrictions, BD-015-01, Research Final Report, State of Florida Department of
Transportation, Tallahassee, FL, 2003.
8 Harwood, D. W., I.B. Potts, D. J Torbic, and W.D. Glauz. Highway/Heavy Vehicle
Interaction. (Commercial Truck and Bus Safety Synthesis Program CTBSSP-SYN-3),
Transportation Research Board, Washington, D.C., 2003.
9 Cate, M. A. and T. Urbanik. Another View of Truck Lane Restrictions. Presented at the
83rd
Annual Meeting of the Transportation Research Board. Transportation Research
Board, Washington, D.C. 2003.
10 Cate, M. A., A. Chatterjee, and V. Kiattokomol. An Evaluation of the Impact of Lane
Use Restrictions for Large Trucks along I-40 Near Knoxville, TNRES 1209/CUT 256,
Research Final Report, State of Tennessee Department of Transportation, 2004.
11 Knipling, R., P. Waller, R.C. Peck, R. Pfefer, T.R. Neuman, K.L. Slack, and K.K. Hardy.
NCHRP Report 500: Guidance for Implementation of the AASHTO Strategic Highway
Safety Plan, Volume 13: A Guide for Reducing Collisions Involving Heavy Trucks.
Transportation Research Board, Washington, D. C., 2004.
12 Hanscom, F.R. (1990). Operational Effectiveness of Truck lane Restrictions ,
Transportation Research Record 1281, TRB, National Research Council, Washington,
D.C., 1990, pp 119-126.
13 Wilmot, C. and M. Khanal. Effect of Speed Limits on Speed and Safety: A Review,
Transport Review, Vol. 19, No. 4, p 315-329, 1999.
14 Sunbelt Research Corporation. Attitudes, Opinions, and Characteristics of Louisiana
Drivers Concerning the 55 mph Speed Limit and Other Highway Safety Issues, Report
37
No: HS-030 613. Louisiana Department of Public Safety, Baton Rouge, Louisiana,
1980.Rajbhandari, R. and J. Daniel. Impacts of the 65 mph Speed Limit on Truck Safety.
Presented at 82nd
Annual Meeting of the Transportation Research Board. Transportation
Research Board, Washington D. C., 2002.
15 Monsere, C. M., C. Newgard, J. Dill, A. Rufolo, E. Wemple, R.L. Bertini, and C.
Milliken. Impacts and Issues Related to Proposed Changes in Oregon’s Interstate Speed
Limits, Portland State University, Portland, Oregon. 2004.
16 Garber, N., J. Miller, B. Yuan, and X. Sun. Safety Effects of Differential Speed Limits
on Rural Interstate Highways, Transportation Research Record 1830, TRB, National
Research Council, Washington D.C., 2003, pp 56-62.
17 Milliken, J. G., Council, F. M., Gainer, T. W., Garber, N. J., Gebbie, K. M., Hall, J. W.,
Lave, C. A., Mason, J. M., Mosteller, F., Nichols, S. D., Oster, C. V., Retting, R. A.,
Sheridan, T. B., Taylor, W. C., Tsebelis, G., Viano, D. C., and Weaver, R. P. Special
Report 254: Managing Speed: Review of Current Practice for Setting and Enforcing
Speed Limits. Transportation Research Board, National Academy Press, Washington, D.
C., 1998.
18 Kweon, Y. J. and K.M. Kockelman. Safety Effects of Speed Limit Changes: Use of
Panel Models, Including Speed, Use, and Design Variables. Transportation Research
Record 1908, TRB, National Research Council, Washington D.C, 2005, pp. 148-158.
19 Chitturi, M.V. and. R.F. Benekohal. Passenger car equivalents for heavy vehicles in
Work Zones. Presented at the 87th
Annual Meeting of the Transportation Research Board.
Transportation Research Board, Washington, D.C. 2007.
20 Ahmed, A., H. Fred and E. Reisman. Developing passenger car equivalents for heavy
vehicles on congested freeways: A capacity based approach. Presented at the 80th
Annual
Meeting of the Transportation Research Board. Transportation Research Board,
Washington, D.C. 2001.
21 Fellendorf, M. and P. Vortisch. Validation of the Microscopic Traffic Flow Model
VISSIM in Different Real-World Situations. Presented at the 80th
Annual Meeting of the
Transportation Research Board. Transportation Research Board, Washington, D.C. 2001.
22 Nicholas J. G., Qun Liu. Identifying the impact of Truck-lane restriction strategies on
safety using simulation.
23 Werner B. and A. Bressler, Traffic Flow on freeway upgrades.
24 Lily E, D. Torbic and Nathan .W. Development of Passenger Car Equivalents for
Freeways ,Two - lane Highways and Arterials.
38
VITA
John Stanley is from Hyderabad, Andhra Pradesh, India. He completed his bachelor’s degree in
civil engineering in June 2004 from University of Kerala, Kerala, India. He joined the master’s
program in civil engineering at Louisiana State University in August 2007. In May 2009, He
will receive the degree of Master of Science in Civil Engineering under the guidance of Dr.
Sherif Ishak.